ENVIRONMENTAL QUALITY BENCHMARKS FOR AQUATIC ECOSYSTEM PROTECTION: DERIVATION AND APPLICATION

Long-term laboratory culture causes contrasting shiftsin tolerance to two marine pollutants in copepodsof the genus Tigriopus

Patrick Y. Sun1&Helen B. Foley1 & Leslie Wu1

& Charlene Nguyen1& Shiven Chaudhry1 &

Vivien W. W. Bao2 & Kenneth M. Y. Leung2,3 & Suzanne Edmands1

Received: 16 May 2017 /Accepted: 2 October 2017 /Published online: 10 October 2017# Springer-Verlag GmbH Germany 2017

Abstract Organismal chemical tolerance is often used to as-sess ecological risk and monitor water quality, yet tolerancecan differ between field- and lab-raised organisms. In thisstudy, we examined how tolerance to copper (Cu) and tribu-tyltin oxide (TBTO) in two species of marine copepods,Tigriopus japonicus and T. californicus, changed across gen-erations under benign laboratory culture (in the absence ofpre-exposure to chemicals). Both copepod species exhibitedsimilar chemical-specific changes in tolerance, with laborato-ry maintenance resulting in increased Cu tolerance and de-creased TBTO tolerance. To assess potential factors underly-ing these patterns, chemical tolerance was measured in con-junction with candidate environmental variables (temperature,UV radiation, diet type, and starvation). The largest chemical-specific effect was found for starvation, which decreasedTBTO tolerance but had no effect on Cu tolerance.Understanding how chemical-specific tolerance can changein the laboratory will be critical in strengthening bioassaysand their applications for environmental protection and chem-ical management.

Keywords Bioassays . Copper . Tributyltin .

Multi-generational . Environmental stress . Antifoulingadditive

Introduction

Toxicological bioassays are an important tool used to establishregulatory practices and ecological risk assessments, and areused internationally as a standard for toxicity testing (Keddyet al. 1995; Power and Boumphrey 2004; Pandard et al. 2006).Bioassays usually employ laboratory-cultured organisms tomake inferences on how natural populations would respond.Laboratory-cultured organisms offer many advantages such asconsistent availability, standardizing sample population acrossmultiple projects, and elimination of environmental acclima-tization effects. However, there is potential for differences inchemical tolerance to develop between field and laboratorypopulations (Berthet et al. 2011). As a result, extrapolatingexperimental findings from the laboratory to the field maynot be appropriate in all settings (Nowak et al. 2008; Woodset al. 1989). Although risk assessments do incorporate safetyfactors to account for uncertainties, how these safety factorsare determined is hotly debated (Konietzka et al. 2014). Theaims of this study are to measure changes in chemical toler-ance across generations in laboratory conditions, to assesscandidate factors responsible for those changes, and to con-tribute to the application of more appropriate safety factorsthat will account for dynamic chemical tolerances. This isimportant because lab-induced change in chemical toleranceswill lead to inappropriate extrapolation of effect thresholds(e.g., predicted no effect concentration, PNEC, or lethal me-dian concentration, LC50). On one hand, this could lead tocostly environmental damage if the laboratory organisms weremore tolerant resulting in underestimated laboratory-derived

Responsible editor: Philippe Garrigues

* Patrick Y. Sunpatricys@usc.edu

1 Department of Biological Science and Wrigley Institute forEnvironmental Studies, University of Southern California, 3616Trousdale PKWY STE 107, Los Angeles, CA 90089, USA

2 The Swire Institute of Marine Science and School of BiologicalSciences, The University of Hong Kong, Pokfulam, Hong Kong,China

3 State Key Laboratory in Marine Pollution, City University of HongKong, Tat Chee Avenue, Kowloon, Hong Kong, China

Environ Sci Pollut Res (2018) 25:3183–3192https://doi.org/10.1007/s11356-017-0398-z

PNEC. On the other hand, if the laboratory organisms used toderive PNEC were more sensitive, then that would lead to anoverestimation and unnecessarily higher cost for pollutioncontrol and cleanup.

Stress tolerance has been shown to shift in the laboratorydue to either physiological processes such as acclimation orgenetic processes such as adaptation. A decrease in stresstolerance could be attributed to a loss of acclimations gainedwhile in the field. For example, in the absence of natural levelsof zinc, Daphnia clones gradually lost their zinc tolerance inthe laboratory (Muyssen et al. 2002). Decreased tolerance hasalso been attributed to loss of genetic diversity. Low geneticdiversity, which laboratory populations often face due to smallpopulation size and inbreeding, has been shown to decreasestarvation tolerance (Stewart et al. 1982) as well as resistanceto contamination (Barata et al. 2000). Tolerance can also bereduced by evolutionary tradeoffs, where selection improvesone trait at the cost of another (Stearns 1989). As an illustra-tion, a laboratory population of Drosophila that evolved ear-lier fertility had a tradeoff of lower desiccation tolerance(Hoffmann et al. 2001). These examples illustrate that therecan be significant loss of tolerance in laboratory populations.

However, stress tolerance can also increase in the laborato-ry. Physiological acclimation has been shown to increase tol-erance following a pre-exposure to sublethal levels of tracemetals (LeBlanc 1982; Kwok et al. 2009; Sun et al. 2014)and salinity (Wu et al. 2014). Likewise, genetic adaptationcan increase tolerance through multiple generation of expo-sure to stressors such as temperature (Kelly et al. 2012).Further, unexpected increases in stress tolerance can also oc-cur in a laboratory setting through indirect selection. One ex-ample is seen when Drosophila raised in ambient temperature(25 °C) developed higher cold tolerance (Condon et al. 2015).Yet, the rate of change in chemical tolerance and the condi-tions responsible are not well characterized. Tolerance canchange gradually over the course of several generations orabruptly after a short period in laboratory culture, driven bythe differences between the field and laboratory conditions.

This study used two species of marine copepods, Tigriopusjaponicus and T. californicus, which both inhabit supralittoraltide pools. Tolerance to two antifouling paint additives, copper(Cu) and tributyltin oxide (TBTO), was measured. AlthoughTBTO has been banned for marine uses (IMO 2001), it wasextensively used before its ban and is still found in marineenvironments through terrestrial runoff and legacy contamina-tion (Díez et al. 2002; Burton et al. 2005; Santos et al. 2010) aswell as illegal uses (Ho et al. 2016). Since the ban on TBTO,Cu has taken its place as the most common antifouling addi-tive. Additionally, these two chemicals represent two broadcategories of marine pollutants. TBTO is a man-made lipo-philic compound that is highly toxic even at low concentra-tions (Champ and Seligman 1996), whereas Cu is a naturallyoccurring trace metal, but especially toxic to marine

invertebrates at higher concentrations (Philips and Rainbow1993). Another point of difference between these twochemicals is that these Tigriopus species have been shown toacclimate to Cu and develop Cu tolerance in the laboratory(Kwok et al. 2009, Sun et al. 2014), while they consistentlyexhibited relatively higher sensitivity to TBTO with noacclimation-like response (Bao et al. 2011; Sun et al. 2014).

This study tracked chemical tolerance across three genera-tions in the laboratory, immediately following collection fromthe field. The goal was to determine how chemical tolerancecan change across several generations under benign laboratoryconditions (culturing was done in the absence of chemicalexposure). To assess factors that might alter tolerance in lab-raised animals, chemical tolerance was also measured in com-bination with several environmental factors that may differbetween field and laboratory environments. Sublethal levelsof environmental stressors were chosen to test the threshold ofphysiological response in these copepods. It was hypothesizedthat these additional stressors would reduce chemical toler-ance by siphoning off resources used in chemical defense(Calow 1991). Ultraviolet irradiation is one of the testedstressors because it is believed to be a constant stressor basedon the general topography of tide pools which offer little to noshelter against direct exposure. Temperature stress is anothercommon stressor in high-tide pool environments, yet closelyregulated in the laboratory. Two types of temperature regi-mens were applied to mimic constant vs. variable heat stress.Diet types were chosen to contrast a standard laboratory dietof dried food with a more natural diet of live algae. Finally, astarvation treatment was tested to mimic conditions likely tooccur in the field.

Methods

Specimen collection and maintenance

Four Tigriopus japonicus populations were collected fromHong Kong in November 2011 and transported to theUniversity of Southern California, USA. The sites of collec-tion in Hong Kong were Gold Coast (22° 22′ 47′′, 113° 58′ 48′′), Shek O (22° 13′ 48′′, 114° 15′ 36′′), Stanley (22° 13′ 11′′,114° 13′ 11′′), and Cape D’Aguilar (22° 12′ 36′′, 114° 15′ 36′′)(CDA). Five populations of Tigriopus californicus were col-lected from California, USA, in November 2011. The collec-tion sites were Santa Cruz (36° 57′ 0′′, − 122° 2′ 59′′), SantaCruz Island (34° 1′ 12′′, − 119° 40′ 48′′), Leo Carrillo(34° 2′ 23′′, − 118° 56′ 23′′), Laguna Beach (33° 32′ 23′′,− 117° 47′ 24′′) (LB), and San Diego (32° 45′ 0′ ′,− 117° 15′ 36′′) (SD). Populations were acclimated to thelaboratory for 1 week before toxicity testing.

All populations were maintained in 37-μm filteredautoclaved seawater (FASW) at 20 °C with a 12-h light/12-h

3184 Environ Sci Pollut Res (2018) 25:3183–3192

dark photoperiod for the duration of the study. Seawater wascollected from theWrigleyMarine Science Center on CatalinaIsland, which is situated within a marine reserve. Salinity wasadjusted as needed to approximately 35 ppt with deionizedwater. After autoclaving, seawater was cooled overnight inan air-permeable container. Cu and TBTO concentrations inseawater were tested (Sun 2016) and levels were found to belower than the allowable threshold stated in the NationalRecommended Water Quality Criteria set by the UnitedStates Environmental Protection Agency. Cu concentrationswere similar to previous reports of coastal seawater from thisregion, the Southern California Bight (Smail et al. 2012).TBTO was no t de t ec t ed ( t e s t ing done by ALSEnvironmental, USA). Cultures were maintained in Nalgenecontainers at a volume of 400 mL with a complete solutionrenewal and feeding performed once a week. Cultures werefed a dry diet, which consisted of 0.04 g ground TetraMin fishfood (Tetra Holding, Inc., USA) and 0.04 g ground Spirulinacyanobacteria (Nutraceutical Science Institute, USA). Thesefood rations are sufficient to be considered ad libitum. Eachnew generation started with three containers per populationwith 100 pre-copulatory pairs each for a total of 300 pairsper population. A generation is defined as the time it takes apre-copulatory pair (a male clasping a virgin female) to pro-duce sexually mature offspring. In the experiment, the esti-mated generation time is 42 days, which is slightly higher thanprevious estimates because pre-copulatory pairs are formedwith sexually immature females and require additional timeto reach maturity and produce offspring. At 14 days post es-tablishment of a new generation, all adult copepods were re-moved to maintain discrete generations. This husbandry pro-tocol has been successfully used to maintain multiple copepodpopulations (Sun et al. 2015) and multiple consecutive gener-ations (Sun et al. 2014).

Multi-generational population tolerance survey

The standard protocol for the multi-generational populationtolerance survey consisted of a contaminant exposure for96 h without feeding or solution renewal in FASW. The envi-ronmental chamber (Thermo Scientific Precision Incubator)was set to a 12-h light/12-h dark photoperiod and held at aconstant 20 °C. Only male copepods were used because malesare generally less stress tolerant than females (Willett 2010;Kelly et al. 2012), and because female stress tolerance mayfluctuate due to maternal transfer during egg development(Raisuddin et al. 2007). Maternal transfer can either increasea female’s susceptibility to stress following the release of eggsor increase their tolerance as they pass on a portion of thechemical stressor to the egg sac relieving their own bodyburden.

Toxicity assays had three replicates per chemical concen-tration. Each replicate had 10 male animals. Each replicate

consisted of a 30-mL petri dish with an approximate volumeof 15 mL. Mortality was assigned to copepods that failed torespond to physical stimulation (Bao et al. 2011). Initial ob-servations were made on a light table and confirmed with adissecting microscope. The range of Cu and TBTO concen-trations used was based on preliminary estimates of medianlethal concentration (LC50), which were obtained from pilotrange-finding tests. Each LC50 used seven concentrations, in-cluding a control, but the actual concentrations were adjustedeach generation for each population due to changing chemicaltolerance. The generation 0 (F0) survey began in November2011, immediately after collection and each consecutive gen-eration survey followed approximately 42 days after the pre-ceding generation. The first acute toxicity assay was donewithin a week of collection.

The Cu stock solution was made with CuSO4·5H2O(Sigma) in nanopure water to a concentration of 1 g L−1 andserially diluted to obtain the target concentration. The TBTOstock solution was made by diluting bis(tributyltin) oxide(EMD, USA) with acetone (purity > 99.5%, MacronChemicals, USA) to a concentration of 0.1 g L−1. The amountof acetone in all treatment solutions including the acetonecontrol for TBTO toxicity assays was ≤ 1.2 × 10−6 v/v. Thisstudy aimed to test if multiple generations of laboratory cul-ture would alter chemical tolerance and to assess effects ofenvironmental variables that may be different between fieldand laboratory environments. It was not intended to generatetoxicity endpoints for regulatory uses. Even though nominalchemical concentrations are not ideal, the practicality of veri-fying each exposure concentration of our multi-species, multi-population, and multi-generation assays was prohibitive.Nominal concentrations were considered adequate for the pur-pose of testing overall temporal patterns of chemicaltolerance.

Tolerance of each population was measured by LC50 cal-culated by the Trimmed Spearman–Karber method (TSKM).Generalized linear models (GLM) were also used because oftheir effectiveness to determine LC50 values by inclusion ofextreme response probabilities of 0 and 1 (Kerr and Meador1996). TBTO LC50 values were calculated using TSKM(Trimmed Spearman–Karber Program v. 1.5, USEnvironmental Protection Agency). While TSKM is an oldermethod, it fares well in comparison to more recent ap-proaches. A study by Hamilton (1980) showed that the confi-dence intervals derived by TSKM performed as well as amaximum likelihood estimator (MLE) and under heavy-tailed tolerance distributions (a distribution characterized bylarge values), TSKM was more accurate than MLE. Cu LC50

values were calculated using GLM due to the abundance ofextreme response probabilities. GLM was calculated with abinomial distribution using the R programing language (RCore Development Team 2013) and package “MASS”(Venables and Ripley 2002). Significant differences between

Environ Sci Pollut Res (2018) 25:3183–3192 3185

generations were inferred from non-overlapping confidenceintervals, which has been shown to be a highly conservativeapproach (Wheeler et al. 2006), which is ideal for this studybecause the goal of these experiments was to identify extremechanges in chemical tolerance.

Environmental stress treatments

All environmental stress treatments used SD (San Diego) co-pepods. Due to the requirement of large numbers of specimensfor each environmental stress assay, large long-term freelybreeding cultures of overlapping generations wereestablished. This is in contrast to the smaller discrete genera-tions maintained during the multi-generational population tol-erance survey. These stress treatments were conducted fromJuly 2012 to August 2013. All toxicity assays following en-vironmental stress were conducted using the standard protocollisted above. LC50 values were calculated using the TSKMand significance was determined according to the methodoutlined in Wheeler et al. (2006).

Ultraviolet irradiation treatment

UVassays were done using the SD population after approxi-mately 21 generations under lab conditions. Males from theSD population were placed in an open Nalgene container witha FASW volume of 400 mL exposed to UV-B at 305 nm for2 h. The 2-h time point was chosen based on field data show-ing that the peak duration of many environmental stressors,such as peak UVor maximum daily temperature, was approx-imately 2 h (unpublished data). Food was withheld during thetreatment. UV exposure was administered with a UV lampwith an intensity of 0.9 mW/cm2 (Cole-Parmer, 6 W/115 V/60 Hz/0.16 A). A control group of SD males was kept inidentical conditions, but without UV. The absence of UVwas confirmed with a UVA/B Light Meter (Sper Scientific,USA). No mortality was measured in UV treatments. Therewas a resting period of 1 h between UVexposure and toxicitytesting.

Diet treatments

Diet assays were done after approximately 28 generationsunder laboratory conditions. Copepods were raised in twodifferent diet treatments for one entire generation before tox-icity tests. The dry diet is our standard laboratory diet of 0.04 gof TetraMin and 0.04 g Spirulina per 400mL replicate, and theanimals were fed once a week. The live algae diet consisted of160 mL Platymonas with a cell count of approximately1,000,000/mL and the copepods were fed once a week. Cellcounts were done on a hemocytometer (Bright-Line, USA)and calculated with the Palmer–Maloney cell count formula.Populations were established with 300 pre-copulatory pairs

per biological replicate, with two biological replicates per diet.After 14 days, all adults were removed to limit cannibalism.Once mature, adult males were collected and used for toxicityassays.

Starvation assays

Starvation assays conducted after approximately 11 genera-tions in the lab. Starvation was induced by isolating malesfor 1 week without food in FASW. There were a total of fivereplicates with 200 males per replicate. Cultures were main-tained using the same husbandry protocol as in the populationsurvey with solution renewal once a week.

Temperature treatments

Temperature assays were done after approximately 18 gener-ations in the lab. The first temperature treatment was the cy-cling temperature treatment, which had environmental condi-tions cycle between 20 °C in the dark for 12 h and 28 °C in thelight for 12 h per day for an entire generation. This treatmentwas based on results from Kelly et al. (2012) which showed acyclic temperature treatment with a maximum temperature of28 °C which was the warmest achievable temperature that hadan impact (generated a temporary increase in thermal toler-ance) without resulting in significant mortality. Pre-copulatorypairs were placed in the cyclic treatment, removed 14 dayspost culture establishment and the resulting males were har-vested once fully mature for toxicity tests. Solution renewaland feeding schedule followed the standard protocol. Cultureswere started with 300 pre-copulatory pairs from the SD pop-ulation in Nalgene containers with a FASW volume of400 mL. The cyclic temperature treatment had a total of fivereplicates.

The constant high-temperature stress treatment at 32 °Cconsisted of a daily 12-h light and 12-h dark cycle for 4 days.Preliminary trials with the San Diego population showed thata treatment of 32 °C for 4 days was the highest temperatureand duration combination that resulted in no significant mor-tality. Males were collected from four replicates for this trialbecause at the time of the assay, the final replicate did not haveenough individuals. The control males were maintained at20 °C for 4 days in their own separate containers.

Results

Multi-generational population tolerance survey

Chemical tolerance was shown to vary across the three gen-erations (i.e., from F0 to F2) under laboratory conditions in thetwomarine copepod species, with a general pattern of increasein Cu tolerance over time and decrease in TBTO tolerance

3186 Environ Sci Pollut Res (2018) 25:3183–3192

over time (Fig. 1). Eight of the nine populations (i.e., all butLB) had an increase in Cu tolerance. In five populations (twofor T. californicus and three for T. japonicus), there were sig-nificant consecutive increases in Cu tolerance. In contrast,TBTO tolerance significantly decreased across seven of thenine tested populations (i.e., all but LC and CDA) during theexperimental period. In two T. japonicus populations, therewere decreases in TBTO tolerance in each consecutive gener-ation under the laboratory setting.

Ultraviolet radiation

UV exposure before toxicity assays significantly decreasedtolerance to both chemicals in all replicates, as determinedby non-overlapping confidence intervals (Fig. 2).

Diet assays

Different diet treatments influenced chemical tolerance(Fig. 3). In both of the Cu replicates and one of the TBTOreplicates, copepods that fed a dry diet (TetraMin andSpirulina) exhibited higher tolerance than those that fed livealgae (Platymonas). The same trend was found in the secondTBTO replicate but confidence intervals could not be calcu-lated for one of the treatments due to an excess of extremevalues (no mortality or complete mortality).

Starvation assay

It was shown that withholding food for 1 week had no signif-icant effect on Cu tolerance in any of the replicates, but sig-nificantly decreased TBTO tolerance in four replicates

(Fig. 4). The same trend was found in the fifth TBTO replicatebut confidence intervals for one treatment are missing due toextreme mortality values.

Temperature

A cycling temperature treatment of 12 h:12 h 20 °C:28 °Cover the course of an entire generation produced limited ef-fects on chemical tolerance. For Cu (Fig. 5a), there was a trendtoward higher tolerance under the cycling treatment, but onlyone of the five replicates showed a significant effect. ForTBTO (Fig. 5b), there was also a trend toward higher toler-ance in the cycling treatment, but no significant effects werefound and several replicates had missing confidence intervals.

A constant temperature treatment at 32 °C for 4 days priorto toxicity assays produced mixed results with the majority ofcomparisons resulting in non-significant differences (Fig. 6).The one exception was a Cu replicate in which the high-temperature treatment reduced tolerance.

Discussion

Multi-generational population tolerance survey

This study examined the temporal component of chemicaltolerance changes under typical laboratory conditions, whereenvironmental parameters are held at mild and constant levels.Our previous work has shown differences in chemical toler-ance among populations of Tigriopus (Sun et al. 2015). Otherstudies have shown tolerance of animal decreases as they arepropagated in the laboratory (Stewart et al. 1982; Barata et al.

a b

c d

Fig. 1 Multi-generationalpopulation tolerance survey of Cu(a–b) and TBTO (c–d) LC50 withTigriopus californicus (a, c) andTigriopus japonicus (b, d). Cuvalues are log10 transformed.Non-overlapping confidenceintervals were used to determinesignificant differences betweengeneration 0 (field-collected) andgeneration 1, as well as betweengeneration 1 and generation 2.Upward arrow indicates asignificant increase from onegeneration to the next, downwardarrow indicates a significantdecrease from one generation tothe next, and X indicates nosignificant change betweengenerations

Environ Sci Pollut Res (2018) 25:3183–3192 3187

2000;Muyssen et al. 2002). In the current study, we found thatTBTO tolerance decreased, which is in line with these previ-ous studies. However, in contrast to the standard expectation,Cu tolerance was found to increase under benign laboratoryconditions. The observation of contrasting shifts in chemicaltolerance signifies a chemical-specific response to laboratoryculturing. Additionally, the shift in chemical tolerance is sim-ilar in both species and across multiple divergent populationsindicating that this may be a relatively conserved pattern.

Previous work has shown that with chronic sublethal ex-posure, adaptation to TBTO can occur, but not in Cu (Sunet al. 2014). The changes in chemical tolerance observed un-der benign conditions in the current study are not easily attrib-utable to adaptation, since changes occurred after a singlegeneration in the laboratory. These changes are more likelyproducts of physiologically plastic processes.

The follow-up environmental assays were designed to elicitphysiological responses and were therefore conducted withinthe span of a single generation under largely sublethal condi-tions. Although the environmental assays were done severalgenerations after the initial population tolerance survey, thepurpose of these follow-up assays was not to make directcomparisons but rather to identify potential candidate environ-mental stressors that could impact chemical tolerance. The

a

b

Fig. 2 Consecutive ultraviolet irradiation and contaminant exposureassay. LC50 values for copper (a) and TBTO (b), with error barsshowing 95% CI. LC50 assays were conducted at 20 °C for 96 h onanimals from one of two treatments: (1) control: constant 20 °C or (2)UV: ultraviolet exposure for 2 h prior to toxicity tests. The UV treatmentreduced tolerance to both Cu and TBTO

a

b

Fig. 3 Different diet treatments significantly impact chemical tolerance.LC50 values for copper (a) and TBTO (b), with error bars showing 95%CI. LC50 assays were conducted at 20 °C for 96 h. The dry diet treatmentwas comprised of TetraMin and Spirulina and the live algae diet consistedof Platymonas administered once a week. Treatment groups were raisedfrom birth on these diets before toxicity assays

a

b

Fig. 4 Starvation assay. LC50 values for copper (a) and TBTO (b), witherror bars showing 95% CI. LC50 assays were conducted at 20 °C for96 h on animals from one of two treatments: (1) control: standard foodconditions or (2) starved: food was withheld for 1 week before toxicityassays

3188 Environ Sci Pollut Res (2018) 25:3183–3192

field is a complex mosaic of factors that would be incrediblydifficult to replicate in the laboratory. However, by testing

candidate factors, we can begin to understand the potentialfor individual stressors to contribute to tolerance changes.

Ultraviolet exposure

UV radiation decreased tolerance of the copepods to bothchemicals. UV was chosen as a candidate stressor becauseit is a particularly strong and consistent influence in tidepool environments where these copepods can be found.Tide pools such as those inhabited by Tigriopus often havelittle to no permanent shade, which allows UV to be a dailypresence in these environments. There is ample evidence ofUVexerting a selective force on other organisms in the field(Rhode et al. 2001; Häder et al. 2007; Cockell andBlaustein 2013). Our results show that exposure to UVsignificantly reduced tolerance to both chemicals. Finiteresources may be used to defend against UV in the field,which could in turn be reallocated in laboratory populationsto mount a greater defense against a toxicity challenge.There seems to be a common mode of toxicity, namelyoxidative stress, between UV and Cu and as a result, a po-tential for defense mechanisms to overlap is reasonable. Anantioxidant, astaxanthin, found in copepods has beenshown to be directly correlated with Cu tolerance as wellas UV-B tolerance (Caramujo et al. 2012). However, be-cause UV decreased tolerance to both chemicals, it doesnot seem likely to explain the chemical-specific shift ob-served in these copepods.

a

b

Fig. 5 Cycling temperatureassay. LC50 values for copper (a)and TBTO (b), with error barsshowing 95% CI. LC50 assayswere conducted at 20 °C for 96 hon animals that had beenmaintained for an entiregeneration under one of twoconditions: (1) control:stable 20 °C or (2) cyclingtemperature: a cyclic temperaturetreatment 20 °C:28 °C for12 h:12 h daily

a

b

Fig. 6 Constant high-temperature stress assay. LC50 values for copper(a) and TBTO (b), with error bars showing 95% CI. LC50 assays wereconducted at 20 °C for 96 h on animals from one of two treatments: (1)control: 20 °C or (2) temp: 32° for 4 h

Environ Sci Pollut Res (2018) 25:3183–3192 3189

Diet treatments

The two diets were chosen based on their potential to representtwo different types of diet regimes. The dried diet was com-posed roughly of dried fishmeal and dried cyanobacteria ex-tract, which is a common dietary regime in the laboratory. Incontrast, the live diet was a live culture of algae which isbelieved to be a much closer match to their natural diet (al-though no formal studies have been conducted what theseparticular copepods actually eat in the field). This is becausethese copepods inhabit splash pools that are irregularly inun-dated and any detritus that would simulate the dried dietwould be an inconsistent nutrient source. On the other hand,photosynthetic microorganisms such as those delivered in thelive algae diet are expected to be a constant nutrient source.Studies show a consistent presence of photosynthetic organ-isms in tide pools through the measurement of pH (Morris andTaylor 1983). During daylight hours, the pH is more basicbecause photosynthesis is offsetting CO2 release from respi-ration, while during night hours, respiration continues withoutthe abatement leading to an increase in pH.

Copepods raised on the dry diet had higher chemical toler-ance than copepods raised on live algae. In other studies, dif-ferent diets have been shown to alter amounts of astaxanthin(Caramujo et al. 2012) and other antioxidants (Gaetke andChow 2003), which are known to influence Cu tolerance(Caramujo et al. 2012). However, preliminary data onastaxanthin concentrations in copepods raised on the two dietsshowed that copepods raised on live algae had higher concen-trations of that carotenoid (unpublished data). With the drieddiet, copepods could obtain higher concentrations of otherantioxidants. The benefits of the dried diet could be due to agreater diversity in components. The dry diet contains animalprotein (TetraMin contains fishmeal) as well as the photosyn-thetic cyanobacteria Spirulina, known as a source antioxidantincluding phycocyanin (Deng and Chow 2010).

Starvation assays

Results showed that a short period of starvation had no influ-ence on Cu tolerance, but resulted in a substantial decrease inTBTO tolerance. This indicates that the reduction in TBTOtolerance in the laboratory could be attributed to dietary prob-lems, despite our attempt to provide abundant food. Thiswould be surprising for several reasons. First, Tigriopus isknown to have extreme starvation resistance and can surviveup to 90 days in autoclaved seawater (unpublished data), sug-gesting the presence of long-term energy stores including fatdeposits such as those found in pelagic copepods (Lee et al.2006). Second, Cu tolerance is known to be energetically de-manding (Lukasik and Laskowski 2007), suggesting that the1-week starvation treatment did not significantly reduce allenergetically demanding biological functions. Third, there is

little evidence that our standard feeding conditions lead todietary problems, as we regularly maintain Tigriopus culturesfor years using the same feeding conditions (e.g., Pritchardet al. 2013). Nevertheless, our results show that dietary restric-tion had a particularly negative impact on TBTO tolerance.This may be because TBTO mainly bioaccumulates in fat dueto its lipophilic properties. Additionally, lipid stores can beadversely affected in the laboratory-cultured copepods by oth-er factors such as body size even under replete conditions(Huho et al. 2007). Thus, under starved conditions, when anorganism begins to metabolize its fat stores for energy, there isan increased exposure of more sensitive tissues to TBTO be-cause there is less lipid to sequester TBTO. Additionally, me-tabolizing fat stores could release any inert TBTO stored infat, thereby increasing exposure to TBTO and causing toxic-ity. The starvation effect may be further magnified by the factthat TBTO can also inhibit feeding and food assimilation rates(Mouneyrac et al. 2011). While our feeding strategy is suffi-cient to propagate laboratory cultures for years, these labora-tory populations may have lower resource reserves than thosein field populations. Ultimately, these results indicate that foodavailability has a chemical-specific impact on contaminanttolerance.

Temperature assays

The cycling temperature treatment produced trends towardhigher chemical tolerance, suggesting a slight hardening ef-fect. Similar hardening effects have been previously docu-mented in Drosophila (Sejerkilde et al. 2003; Loeschcke andHoffmann 2007). A temperature of 28 °C may be too low toinduce significant stress. The average maximum summer tem-perature experienced in the SD habitat has recently beenshown to be well above 28 °C (Leong et al. in press). Heatstress may have minimal effect on SD fitness at all tempera-tures below a small lethal window. Many intertidal organismsare already living at or close to their thermal limit (Somero2010) and may have evolved the ability to maintain homeo-stasis until a tipping point very close to their lethal tempera-ture. This would result in a very small thermal or temporalwindow to observe sublethal effects.

Constant heat stress for 4 days produced few changes inchemical tolerance. There is ample evidence that high temper-ature has exerted strong selective pressure on Tigriopus pop-ulations across its evolutionary history. T. californicus hasbeen shown to be locally adapted to high-temperature stressand exhibits a strong cline in temperature tolerance (Willett2010; Kelly et al. 2012). Likewise, it was discovered that inthese copepods, there is a positive correlation between Cutolerance and air temperature at collection site (Sun et al.2015). However, these results show that constant heat stresshad minimal effects on chemical tolerance. This again, may be

3190 Environ Sci Pollut Res (2018) 25:3183–3192

due to a very small thermal window where sublethal effectscan be observed.

Conclusion

Tolerance to two common marine pollutants has been shownto change under benign laboratory conditions for two copepodspecies. Additionally, the direction of change is chemical spe-cific, with Cu tolerance increasing and TBTO tolerance de-creasing. This contradicts the expectation that organismal fit-ness will generally decline in the laboratory. A search forcandidates to explain these changes revealed that chemicaltolerance is significantly impacted by several environmentalparameters. UV was shown to lower chemical tolerance whilethe standard laboratory diet was shown to increase chemicaltolerance. These results suggest that laboratory populations,which are protected from UVand fed a protective diet, likelyhave certain chemical tolerances that are higher than their wildcounterparts. This result, coupled with the observed increasein Cu tolerance under benign laboratory conditions, suggeststhat tolerance measurements derived from laboratory studiesmay drastically overestimate tolerance of wild populations inthe field.

The exact mechanism responsible for the contrasting shiftsin chemical tolerance remains elusive, but the present resultsdemonstrate that chemical tolerance is highly labile.Understanding the pattern as well as the underlying cause ofthese changes will be critical for effective implementation ofbioassays as a tool for environmental protection and chemicalmanagement.

Funding This study was supported by Sea Grant (NA10OAR417005) toSE and partially supported by the Research Grants Council of the HongKong SAR Government through a General Research Fund(HKU703511P) to KMYL.

References

Bao VWW, Leung KMY, Qiu JW, LamMHW (2011) Acute toxicities offive commonly used antifouling booster biocides to selected sub-tropical and cosmopolitan marine species. Mar Pollut Bull 62:1147–1151

Barata C, Baird DJ, Amat F, Soares AMVM (2000) Comparing popula-tion response to contaminants between laboratory and field: an ap-proach using Daphnia magna ephippial egg banks. Funct Ecol 14:513–523. https://doi.org/10.1046/j.1365-2435.2000.00445.x

Berthet B, Leung KMY, Amiard-Triquet C (2011) Inter- and intraspecificvariability of tolerance: implications for bioassays and biomonitor-ing. In: Amiard-Triquet C, Rainbow PS, Romeo M (eds) Toleranceto environmental contaminants. CRC Press, Boca Raton, pp 49–84

Burton ED, Phillips IR, Hawker DW (2005) In-situ partitioning ofbutyltin compounds in estuarine sediments. Chemosphere 59:585–592. https://doi.org/10.1016/j.chemosphere.2004.10.067

Calow P (1991) Physiological costs of combating chemical toxicants:ecological implications. Comp Biochem Physiol Part C: CompPharmacol 100:3–6. https://doi.org/10.1016/0742-8413(91)90110-F

Caramujo M-J, de Carvalho CCCR, Silva SJ, Carman KR (2012) Dietarycarotenoids regulate astaxanthin content of copepods and modulatetheir susceptibility to UV light and copper toxicity. Mar Drugs 10:998–1018. https://doi.org/10.3390/md10050998

Champ MA, Seligman PF (eds) (1996) Organotin: environmental fateand effects. Chapman & Hall, London, p 623

Cockell C, Blaustein AR (2013) Ecosystems, evolution, and ultravioletradiation. Springer Science & Business Media

Condon C, Acharya A, Adrian GJ, Hurliman AM, Malekooti D, NguyenP, Zelic MH, Angilletta MJ (2015) Indirect selection of thermaltolerance during experimental evolution of Drosophilamelanogaster. Ecol Evol 5:1873–1880. https://doi.org/10.1002/ece3.1472

Deng R, Chow T-J (2010) Hypolipidemic, antioxidant andantiinflammatory activities of microalgae Spirulina. CardiovascTher 28:e33–e45. https://doi.org/10.1111/j.1755-5922.2010.00200.x

Díez S, Abalos M, Bayona JM (2002) Organotin contamination in sedi-ments from the Western Mediterranean enclosures following 10years of TBT regulation. Water Res 36:905–918

Gaetke LM, Chow CK (2003) Copper toxicity, oxidative stress, and an-tioxidant nutrients. Toxicol Environ 189:147–163. https://doi.org/10.1016/S0300-483X(03)00159-8

Hamilton MA (1980) Inference about the ED50 using the trimmedSpearman-Karber procedure—a Monte Carlo investigation.Communications in Statistics—Simulation and Computation 9:235–254. doi:https://doi.org/10.1080/03610918008812152

Häder D-P, Kumar HD, Smith RC, Worrest RC (2007) Effects of solarUV radiation on aquatic ecosystems and interactions with climatechange. Photochem Photobiol Sci 6:267–285. https://doi.org/10.1039/b700020k

Ho KKY, Zhou GJ, Xu EGB, Wang XH, Leung KMY (2016) Long-termspatio-temporal trends of organotin contaminations in the marineenvironment of Hong Kong. PLoS One 11(5):e0155632. https://doi.org/10.1371/journal.pone.0155632

Hoffmann AA, Hallas R, Sinclair C, Partridge L (2001) Rapid loss ofstress resistance in Drosophila melanogaster under adaptation tolaboratory culture. Evolution 55:436–438. https://doi.org/10.1554/0014-3820(2001)055[0436:RLOSRI]2.0.CO;2

Huho BJ, Ng’habi KR, Killeen GF, Nkwengulila G, Knols BGJ,Ferguson HM (2007) Nature beats nurture: a case study of the phys-iological fitness of free-living and laboratory-reared male Anophelesgambiae s.l. J Exp Biol 210:2939–2947. https://doi.org/10.1242/jeb.005033

Keddy CJ, Greene JC, Bonnell MA (1995) Review of whole-organismbioassays: soil, freshwater sediment, and freshwater assessment inCanada. Ecotoxicol Environ Saf 30:221–251. https://doi.org/10.1006/eesa.1995.1027

Kelly MW, Sanford E, Grosberg RK, (2012) Limited potential for adap-tation to climate change in a broadly distributed marine crustacean.Proc R Soc B rspb20110542. doi:https://doi.org/10.1098/rspb.2011.0542

Kerr DR, Meador JP (1996) Modeling dose response using generalizedlinear models. Environ Toxicol and Chem 15:395–401. https://doi.org/10.1002/etc.5620150325

Konietzka R, Schneider K, Ritter L (2014) Extrapolation factors andsafety factors in toxicology. In: Reichl F-X, Schwenk M (eds)Regul. Toxicol. Springer Berlin Heidelberg, pp 431–438. https://doi.org/10.1007/978-3-642-35374-1_59

Kwok KWH, Grist EPM, Leung KMY (2009) Acclimation effect andfitness cost of copper resistance in the marine copepod Tigriopusjaponicus. Ecotoxicol Environ Saf 72:358–364. https://doi.org/10.1016/j.ecoenv.2008.03.014

Environ Sci Pollut Res (2018) 25:3183–3192 3191

LeBlanc GA (1982) Laboratory investigation into the development ofresistance of Daphnia magna (strauus) to environmental pollutants.Environ Pollut Series A, Ecol Biol 27:309–322. https://doi.org/10.1016/0143-1471(82)90159-3

Lee RF, HagenW, Kattner G (2006) Lipid storage in marine zooplankton.Mar Ecol Prog Ser 307:273–306. https://doi.org/10.3354/meps307273

LeongW, Sun PY, Edmands S (in press) Latitudinal clines in temperatureand salinity tolerance in tidepool copepods. J Hered. https://doi.org/10.1093/jhered/esx061

LoeschckeV, HoffmannAA (2007) Consequences of heat hardening on afield fitness component in Drosophila depend on environmentaltemperature. Am Nat 169:175–183. https://doi.org/10.1086/510632

Lukasik P, Laskowski R (2007) Increased respiration rate as a result ofadaptation to copper in confused flour beetle, Tribolium confusumJacquelin du Val. Bull Environ Contam Toxicol 79:311–314. https://doi.org/10.1007/s00128-007-9227-2

Mouneyrac C, Leung PTY, Leung KMY (2011) Cost of tolerance. In:Amiard-Triquet C, Rainbow PS, Romeo M (eds) Tolerance to envi-ronmental contaminants. CRC Press, Boca Raton, pp 265–297

Morris S, Taylor AC (1983) Diurnal and seasonal variation in physico-chemical conditions within intertidal rock pools. Estuar Coast ShelfSci 17:339–355. https://doi.org/10.1016/0272-7714(83)90026-4

Muyssen BTA, Janssen CR, Bossuyt BTA (2002) Tolerance and acclima-tion to zinc of field-collected Daphnia magna populations. AquatToxicol 56:69–79. https://doi.org/10.1016/S0166-445X(01)00206-5

Nowak C, Czeikowitz A, Vogt C, OetkenM, Streit B, Schwenk K (2008)Variation in sensitivity to cadmium among genetically characterizedlaboratory strains of the midge Chironomus riparius. Chemosphere71:1950–1956. https://doi.org/10.1016/j.chemosphere.2007.12.023

Pandard P, Devillers J, Charissou A-M, Poulsen V, Jourdain M-J, FérardJ-F, Grand C, Bispo A (2006) Selecting a battery of bioassays forecotoxicological characterization of wastes. Sci Total Environ 363:114–125. https://doi.org/10.1016/j.scitotenv.2005.12.016

Philips DJH, Rainbow PS (1993) Biomonitoring of trace aquatic contam-inants. Springer, London 371 p

Pritchard VL, Knutson V, Lee M, Zieba JT, Edmands S (2013) Fitnessandmorphological outcomes of many generation of hybridization inthe copepod Tigriopus californicus. J Evol Biol 26:416–433

Power EA, Boumphrey RS (2004) International trends in bioassay use foreffluent management. Ecotoxicology 13:377–398

R Core Team (2013) R: a language and environment for statistical com-puting. R Foundation for Statistical Computing, Vienna, Austria.ISBN 3-900051-07-0, URL http://www.R-project.org/

Rhode SC, Pawlowski M, Tollrian R (2001) The impact of ultravioletradiation on the vertical distribution of zooplankton of the genusDaphnia. Nature 412:69–72. https://doi.org/10.1038/35083567

Santos DM d, Sant’Anna BS, Sandron DC, Cardoso de Souza S, CristaleJ, Marchi MRR d, Turra A (2010) Occurrence and behavior ofbutyltins in intertidal and shallow subtidal surface sediments of anestuarine beach under different sampling conditions. Estuar CoastShelf Sci 88:322–328. https://doi.org/10.1016/j.ecss.2010.04.009

Sejerkilde M, Sørensen JG, Loeschcke V (2003) Effects of cold and heathardening on thermal resistance in Drosophila melanogaster. J InsectPhysiol 49:719–726

Smail EA, Webb EA, Franks RP, Bruland KW, Sañudo-Wilhelmy SA(2012) Status of metal contamination in surface waters of the coastalocean off Los Angeles, California since the implementation of theClean Water Act. Environ Sci Technol 46:4304–4311. https://doi.org/10.1021/es2023913

Somero GN (2010) The physiology of climate change: how potentials foracclimatization and genetic adaptation will determine ‘winners’ and‘losers’. J Exp Biol 213:912–920. https://doi.org/10.1242/jeb.037473

Stearns SC (1989) Trade-offs in life-history evolution. Funct Ecol 3:259–268. https://doi.org/10.2307/2389364

Stewart NP, Callow LL, Duncalfe F (1982) Biological comparisons be-tween a laboratory-maintained and a recently isolated field strain ofBoophilus microplus. J Parasitol 68:691–694

Sun PY (2016) The evolution of pollution tolerance in the marine cope-pod Tigriopus. Dissertation, University of Southern California, Ph.D

Sun PY, Foley HB, Handschumacher L, Suzuki A, Karamanukyan T,Edmands S (2014) Acclimation and adaptation to common marinepollutants in the copepod Tigriopus californicus. Chemosphere 112:465–471. https://doi.org/10.1016/j.chemosphere.2014.05.023

Sun PY, Foley HB, Bao VWW, Leung KMY, Edmands S (2015)Variation in tolerance to common marine pollutants among differentpopulations in two species of the marine copepod Tigriopus.Environ Sci Pollut Res 22(20):16143–16152

Venables WN, Ripley BD (2002) Modern applied statistics with S, 4thedn. Springer, New York 0-387-95457-0

Wheeler MW, Park RM, Bailer AJ (2006) Comparing median lethal con-centration values using confidence interval overlap or ratio tests.Environ Toxicol Chem 25:1441–1444

Woods PE, Paulauskis JD, Weigt LA, Romano MA, Guttman SI (1989)Genetic variation in laboratory and field populations of the midge,Chironomus tentans fab.: implications for toxicology. EnvironToxicol Chem 8:1067–1074. https://doi.org/10.1002/etc.5620081112

WuC-S, YangW-K, Lee T-H, Gomez-Mestre I, KamY-C (2014) Salinityacclimation enhances salinity tolerance in tadpoles living in brackishwater through increased Na+, K+−ATPase expression. J Exp Zool321:57–64. https://doi.org/10.1002/jez.1837

3192 Environ Sci Pollut Res (2018) 25:3183–3192